Support pydantic dataclasses in structured outputs

[9dogs]

Confirm this is a feature request for the Python library and not the underlying OpenAI API.

• This is a feature request for the Python library

Describe the feature or improvement you're requesting

It would be great to be able to use data schemas defined as pydantic dataclasses in structured outputs. E.g.

from pydantic.dataclasses import dataclass
from openai import OpenAI

client = OpenAI()

@dataclass
class CalendarEvent:
    name: str
    date: str
    participants: list[str]

completion = client.beta.chat.completions.parse(
    model="gpt-4o-2024-08-06",
    messages=[
        {"role": "system", "content": "Extract the event information."},
        {"role": "user", "content": "Alice and Bob are going to a science fair on Friday."},
    ],
    response_format=CalendarEvent,
)

event = completion.choices[0].message.parsed

Pydantic dataclasses can be easily transformed into JSON schema via the model_json_schema function.

Additional context

No response

[UtsavChokshiCNU]

@9dogs ... Are you able to even run the code from their tutorial with pydanitc?
It returns me an error :
AttributeError: 'Completions' object has no attribute 'parse'

[parth126]

@UtsavChokshiCNU You need to update the api version to the latest version. Older version did not have parse attribute

[RobertCraigie]

Support for pydantic.dataclasses.dataclass has been added in v1.42.0